PKP3 promotes RB phosphorylation and E2F1 activity

Figure 32 │ Growth retardation is not a consequence of premature differentiation. A schematic of the epidermis demonstrates the expression of differentiation markers in distinct epidermal layers (upper left panel, created with biorender.com, modified from (Matsui and Amagai, 2015). Desmoglein 1, keratin 1, and keratin 10 are highly expressed in the spinous layer, whereas corneodesmosin, involucrin, and loricrin show increased expression in the upper corneal and granular layers. For analysis of the PKP3-dependent expression pattern of differentiation markers, WT and PKP3-KO cells were grown for 72 h in medium with or without Ca²⁺ and lysed in SDS lysis buffer. Total cell extracts were analyzed by western blotting with the indicated antibodies.

Representative western blots (upper right panel) showing the protein levels of PKP3 and epidermal differentiation markers. Ponceau S staining was used as a loading control. PKP3-KO cells grown in HCM showed unaltered or decreased protein levels of all differentiation markers compared to WT cells. The bar plots (lower panels) depict the protein amounts (+ s.d.; n=3) normalized to Ponceau S staining and relative to WT cells grown in medium without Ca²⁺ (first lane in western blot). Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons test.

Collectively these results show that loss of PKP3 led to an enrichment of cells in G0/G1 phase and a prolonged G1 phase. As a consequence, S phase entry was retarded. These data indicate that PKP3 promotes proliferation and cell cycle progression in murine keratinocytes. In the following chapter, the mechanistic basis of PKP3’s role in the G1-S phase transition will be analyzed.

(CDK)4/CDK6 and CDK2, respectively. These activated complexes promote RB phosphorylation which leads to E2F dissociation. The priming hypo-phosphorylation is mediated by cyclin D-CDK4/6, which mono-phosphorylate RB. Subsequently, cyclin E-CDK2 completes RB phosphorylation and its inactivation. Hyper-phosphorylated RB dissociates from E2F1 and can be translocated into the cytoplasm in an exportin-1 dependent manner (Jiao et al., 2008). The release of E2F1 leads to its full activation and allows the expression of E2F1 target genes that encode proteins necessary for S phase transition.

Figure 33 │ PKP3 regulates G1 phase in murine keratinocytes. (A) A schematic of the key events initiating G1-S phase progression demonstrates RB pathway as a central control mechanism (created with biorender.com, modified from (Giacinti and Giordano, 2006)). Un-phosphorylated, activated RB binds to E2F1 transcription factor, which represses its transcriptional activity and blocks cell cycle progression. When a cell receives mitogenic signals, complex formation of CDK4/6 and CDK2 with their corresponding cyclins is triggered, leading to their full activation. Activated cyclin-CDK complexes partially deactivate RB by

phosphorylation. Subsequently, E2F1 is released and mediates expression of genes that encode proteins necessary for G1-S phase transition. (B) For analysis of the putative effect of PKP3 on the RB pathway, WT, PKP3-KO, and WT+PKP3 cells were grown for 24 h in LCM and lysed in SDS lysis buffer. Total cell extracts were analyzed by western blotting with the indicated antibodies. Representative western blots (left panel) showing the protein level of PKP3 and distinct G1 phase markers. In the PKP3 blot, the lower lane represents endogenous PKP3, the upper lane reflects PKP3-GFP. β-actin was used as a loading control. PKP3-KO cells showed unaltered protein levels of cyclin D1, D2, and E1, as well as CDK4, phospho-CDK4-Thr¹⁷², and E2F1 compared to WT cells. CDK6 protein amount was reduced in PKP3-KO cells, but the level of its active form, phospho-CDK6-Thr¹⁷⁷, was unaltered. CDK2 protein level was increased in PKP3-KO cells and decreased in WT+PKP3 cells, but the amount of its activated form (phospho-CDK2-Thr¹⁶⁰) was unaltered or slightly increased. Importantly, RB and phospho-RB-Ser^807/811 revealed PKP3-dependent levels, as shown by decreased amounts in PKP3-KO cells and increased levels in WT+PKP3 cells. A schematic of CDK activation (lower mid panel, created with biorender.com) demonstrates CDK4/6 and CDK2 phosphorylation sites, with indicated analyzed phosphosites (marked in bold). The bar plots (upper panel and lower right panel) depict the protein amounts (+ s.d.; n=3) normalized to β-actin and relative to WT cells (first lane in western blot). Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons test.

To analyze the putative effect of PKP3 on the RB pathway, the amount and activity of distinct proteins involved in RB regulation were quantified by western blotting (Figure 33B).

In PKP3-KO cells, protein levels of cyclin D1, D2, and E1 were unaltered. Furthermore, CDK4 protein amount and activity (i.e., phosphorylation at Thr¹⁷²) were not affected by the loss of PKP3. However, CDK6 protein level and its phosphorylation at Tyr²⁴, which inhibits cyclin D-CDK6 association and prevents cell cycle progression, were decreased in PKP3-KO cells, suggesting low levels of inactive CDK6. Despite its reduced protein level, activation of CDK6 by phosphorylation at Thr¹⁷⁷ was not affected by PKP3. In contrast, the CDK2 protein level was significantly increased in PKP3-KO cells, but the amount of activated CDK2 (i.e., phosphorylation at Thr¹⁶⁰) was unaltered. WT+PKP3 cells showed decreased CDK2 protein levels without a decrease in its activation. This suggests that PKP3 primarily affects the amount of CDK6 and CDK2 protein levels without significant effects on their activation.

Even though the activities of the cyclin-CDK complexes were essentially unaltered, I investigated whether RB and/or E2F1 activity was affected by PKP3. The most dramatic effect was observed downstream of cyclin-CDKs. Total RB levels were significantly reduced in PKP3-KO and increased in WT+PKP3 cells. To analyze the effect on RB phosphorylation, I used an antibody raised against human phospho-RB-Ser^807/811, a well-known CDK4/CDK6 and CDK2 phosphorylation site (Rubin, 2013; Narasimha et al., 2014).

To validate this antibody in murine keratinocytes, WT cells were treated with Lambda PP, which is active towards phosphorylated serine, threonine, and tyrosine residues (Figure S6). Treatment with Lambda PP did not alter PKP3 or RB levels, but the phospho-RB signal was abolished confirming the antibody’s specificity for phospho-phospho-RB. Not only total RB was affected by PKP3, phospho-RB also revealed PKP3-dependent changes (Figure 33B). PKP3-KO cells showed an even more pronounced reduction in phospho-RB

compared to total RB, whereas WT+PKP3 cells showed elevated levels of phospho-RB.

These data suggest that PKP3 might contribute to the regulation of RB phosphorylation.

PKP3-KO cells were obtained from PKP3-KO mice, which were generated by the Cre/loxP-site specific recombination system and were specified by the depletion of the mouse Pkp3 exon 2-4 (Figure 34A) (Sklyarova et al., 2008). PKP3-KO cells were characterized by the loss of PKP3 protein (Figure 33B) and, surprisingly, by high levels of non-protein coding PKP3 RNA (Figure 34B). To analyze whether this non-protein coding RNA might affect the expression and regulation of components of the RB pathway, WT and PKP3-KO cells were treated with control or PKP3-directed (siPKP3) siRNAs. Depletion of PKP3 resulted in decreased PKP3 mRNA levels in both WT and PKP3-KO cells. Nonetheless, the loss of the non-protein coding PKP3 RNA in PKP3-KO cells did not affect CDK6 and RB mRNA levels, which were reduced to the same extent as in PKP3-KO cells with high levels of the truncated PKP3 RNA. This excludes that the observed effects on cell cycle associated genes were caused by the non-protein coding PKP3 RNA but are due to the loss of the PKP3 protein.

Figure 34 │ Increased non-coding PKP3 RNA levels in PKP3-KO cells does not affect CDK6 and RB mRNA amount. (A) Schematic of Pkp3-WT allele and mutated Pkp3 allele with depletion of exon 2-4 (created with biorender.com, modified from (Sklyarova et al., 2008)). Exons are indicated as numbered blue boxes.

Exon 1 starts with the ATG start codon, exon 13 ends with the TAG stop codon. PKP3-KO cells were obtained from PKP3-KO mice, which were generated by knockout of exon 2-4. (B) WT and PKP3-KO cells were transfected with non-targeting (siCtrl) or PKP3-directed (siPKP3) siRNAs, grown for 72 h in LCM, and processed for qRT-PCR. PKP3-KO cells showed increased levels of non-protein coding RNA, which was reduced to the same extent as in WT cells after PKP3 knockdown (left panel). CDK6 and RB mRNA were reduced in siCtrl and siPKP3 treated PKP3-KO cells compared to WT cells (right panel). The bar plots depict the mRNA amounts (+ s.d.; n=5) normalized to Eif3k as an invariant endogenous control (reference gene) and relative to WT cells. Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons test.

Although PKP3-KO cells showed altered CDK6 and RB protein levels (Figure 33B), PKP3 depletion in WT cells was not sufficient to decrease CDK6 and RB mRNA levels (Figure 34B), which might be due to the short time frame of the knockdown experiment. To analyze this effect in detail, protein stability of PKP3, CDK6, and RB was determined. WT, PKP3-KO, and WT+PKP3 cells were treated with cycloheximide, an inhibitor of protein biosynthesis which prevents translational elongation. After treatment and lysis of the cells, the protein abundance was analyzed by western blotting (Figure 35). The highly unstable protein c-MYC was used as positive control (Welcker et al., 2004; Li et al., 2015). Even after prolonged cycloheximide treatment, protein abundance of CDK6 and RB were only slightly decreased in WT cells, suggesting high protein stability. Highly stable proteins do not need continuous translation from mRNA, which might contribute to smaller effects of PKP3 knockdown in WT cells compared to its depletion in PKP3-KO cells. Moreover, CDK6 and RB protein levels after cycloheximide treatment were unaltered in PKP3-KO and WT+PKP3 cells compared to WT cells. Thus, it seems that PKP3 does not affect protein stability of CDK6 and RB.

Figure 35 │ PKP3 does not affect protein stability of CDK6 and RB. A cycloheximide chase assay was performed to measure steady-state protein stability. WT, PKP3-KO, and WT+PKP3 cells were grown for 24 h in LCM and treated with cycloheximide (CHX, 200 µg/ml) for 8 h or 16 h. Cells were lysed in SDS lysis buffer.

Total cell extracts were analyzed by western blotting with the indicated antibodies. Representative western blots (left panel) showing the protein level of PKP3, CDK6, and RB. In the PKP3 blot, the lower lane represents endogenous PKP3, the upper lane reflects PKP3-GFP. β-actin was used as a loading control. c-MYC was used as positive control. Whereas c-MYC showed dramatically reduced protein amounts in all three cell lines, PKP3, CDK6, and RB protein levels were unaltered even after 16 h cycloheximide. The graph (right panels) depicts the protein abundance (± s.d.; n=3) normalized to β-actin and relative to 0 h cycloheximide. Statistical significance was determined by one-way ANOVA.

As CDK6 can phosphorylate RB and PKP3-KO cells showed diminished protein levels of CDK6 and phospho-RB, I determine whether increased expression of CDK4 or CDK6 can rescue RB phosphorylation in PKP3-KO cells. Thus, HA-tagged CDK4 or CDK6 were expressed in all three cell lines (Figure 36). Neither the ectopic expression of CDK4 nor expression of CDK6 increased RB phosphorylation in PKP3-KO cells. This suggests that the reduced CDK6 protein level in PKP3-KO cells is not the main driver of the prolonged G1 phase.

Figure 36 │ Ectopic expression of CDK4 or CDK6 does not rescue RB phosphorylation in PKP3-KO cells. For analyzing the effect of CDK4 and CDK6 overexpression, WT, PKP3-KO, and WT+PKP3 cells were transfected with the indicated Venus2-HA-CDK constructs or a plasmid without insert (mock) and grown for 24 h in LCM. Cells were lysed in SDS lysis buffer. Total cell extracts were analyzed by western blotting with the indicated antibodies. Representative western blots (left panel) showing the protein level of CDK4/6 and phospho-RB. Ponceau S staining was used as a loading control. Amount of phospho-RB was not increased in PKP3-KO cells after ectopic expression of CDK4 or CDK6. The bar plot (right panel) depicts the level of phospho-RB (+ s.d.; n=3) normalized to Ponceau S staining and relative to mock treatment in WT cells (first lane in western blot). Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons test.

RB phosphorylation is the key event in G1-S phase transition by promoting E2F1 release and regulating its transcriptional activity. E2F1 protein level was not affected by PKP3 (Figure 33B). As PKP3-KO cells showed impaired RB phosphorylation, I hypothesized that E2F1 activity might also be reduced. To analyze the transcriptional activity of E2F1, mRNA levels of selected E2F1 target genes involved in cell cycle control, such as E2F3, CCNA2 (cyclin A2), CCNB1 (cyclin B1), CDK1, minichromosome maintenance complex components 3 and 6 (MCM3, MCM6), thymidylate synthase (TYMS), and flap endonuclease 1 (FEN1), were quantified by qRT-PCR (Figure 37). In PKP3-KO cells, mRNA levels of these E2F1 target genes were decreased. This indicates that the loss of PKP3 resulted in reduced transcriptional activity of E2F1 and was further supported by GSEA-based HALLMARK data (Figure S5), in which downregulated protein coding genes in PKP3-KO cells and upregulated protein coding genes in WT+PKP3 cells were highly associated with E2F targets.

Figure 37 │ PKP3 promotes transcriptional activity of E2F1. For analyzing the transcriptional activity of E2F1, WT and PKP3-KO cells were grown for 24 h in LCM and processed for qRT-PCR. Analysis of mRNA of E2F1 targets involved in cell cycle control revealed decreased levels in PKP3-KO cells compared to WT cells.

The bar plot depicts the log2 mRNA fold changes (+ s.d.; n=7) normalized to Eif3k as an invariant endogenous control (reference gene) and relative to WT cells. Statistical significance was determined by a student’s unpaired two tailed t-test.

Taken together, these data show that PKP3 promotes the G1-S phase transition by increasing RB phosphorylation, thereby enhancing E2F1 activity. In the following chapter, the molecular mechanism underlying the regulation of the G1 to S phase transition by PKP3 will be analyzed.